Load Story Context

Overview

Use this skill to load the current story project as a coherent working context. It should inspect story.yaml, normalize essential metadata, identify missing or inconsistent files, and summarize the minimum reliable context that downstream skills should use.

When To Use

Use this skill when:

• the current directory is expected to be a story project;
• story.yaml exists but may be incomplete or inconsistent;
• multiple project files exist and the agent needs a clean shared context before planning, drafting, or reviewing;
• the user wants to audit project readiness;
• a downstream skill would otherwise need to re-discover basic metadata from scratch.

This skill is especially useful before define-style-bible, create-character-bible, generate-high-level-outline, plan-chapter, draft-chapter, chapter-review, or continuity-check.

Required Inputs

Gather or infer these inputs:

• story.yaml;
• root project files and folders;
• key story files that help confirm current state;
• existing workflow status if present.

Ask questions only when essential metadata is missing and cannot be inferred safely.

Files To Read

Always read:

• story.yaml when present;
• README.md when present.

Read these when present and useful for state detection:

• 00_meta/premise.md
• 00_meta/goals.md
• 00_meta/constraints.md
• 01_style/*
• 02_story/*
• 03_characters/*
• 04_outline/*
• 05_draft/*
• 06_review/*
• 07_continuity/*

Do not load everything blindly if only a narrow downstream task needs a smaller slice. Prefer a targeted scan after validating the shared metadata.

Files To Write

Write or update only when useful:

• story.yaml
• optionally README.md if the project summary is clearly stale and the workflow calls for a refresh

This skill should be conservative. Its main output is normalized metadata and a reliable context summary, not broad content generation.

Metadata Contract

Treat story.yaml as the project's source of truth for shared metadata. At minimum, preserve or normalize:

project:
  title: ""
  format: "novel"
  status: ""

story:
  genre: ""
  subgenre: ""
  language: ""
  audience: ""
  estimated_length: ""
  premise_status: ""

narrative:
  tone: ""
  voice: ""
  pov: ""
  tense: ""

workflow:
  canonical_root: "."
  initialized_at: ""
  next_skill: ""

Normalization rules:

• YAML keys remain in English;
• missing required sections should be created if enough evidence exists to do so safely;
• existing values should be preserved unless clearly invalid or contradicted by stronger project evidence;
• story.language is mandatory for content-writing skills and should be treated as blocking when absent;
• human-readable content and free-text metadata values should use correct orthography and normal Unicode for the selected language;
• workflow.canonical_root should remain "." for this repository convention.

Procedure

1. Inspect the current directory and confirm whether it resembles a story project.
2. Read story.yaml and validate that required top-level sections exist.
3. Detect missing, empty, contradictory, or stale metadata.
4. Cross-check metadata against obvious evidence from core files such as premise, style, outline, or draft files.
5. Normalize story.yaml conservatively:
   • add missing sections or keys;
   • preserve valid existing values;
   • mark uncertain fields as provisional through adjacent status fields or explicit notes when the project pattern supports that.
6. Identify workflow readiness:
   • initialized only;
   • style defined;
   • cast defined;
   • outline ready;
   • chapter planning in progress;
   • drafting underway;
   • review artifacts present;
   • continuity tracking present.
7. Produce a short context summary that downstream skills can rely on.

Output Requirements

The output should summarize:

• whether the current directory is a valid story project;
• current metadata reliability;
• selected story language;
• current workflow stage;
• missing or inconsistent core files;
• recommended next skill.

If story.yaml was changed, state what was normalized.

Quality Bar

The resulting context load should allow another skill to answer:

• what project it is working on;
• which language to write in;
• what stage the project is in;
• which files are authoritative;
• what important context is still missing.

Output

After running the skill, provide:

• a short project status summary;
• normalized metadata fields, if any;
• the most important missing or inconsistent items;
• the recommended next step.

Definition Of Done

This skill is done when:

• the current project has a usable story.yaml;
• required shared metadata is present or clearly flagged as missing;
• the current workflow stage is understandable;
• downstream skills can proceed with less ambiguity than before.

Restrictions

• Do not rewrite creative content just to make metadata look tidy.
• Do not invent missing facts without evidence.
• Do not silently change the project's selected language.
• Do not treat every missing optional file as a blocker.

Example Invocation

Use $load-story-context to inspect the current story project, normalize story.yaml, and tell me which skill should run next.